Resolve artifact ambiguities using shortest match

This changes the behavior of `select_platform()` to resolve ambiguities
using a two-step process.  Previously, we would simply sort the
resultant triplets and return the last one (so as to provide asemi-
arbitrary stability, and also to prefer higher version numbers over
lower version numbers in tags).  However, with the proliferation of tags
(and especially the new `sanitize=memory` tags) we need a way to exclude
these "extra" tags when our `HostPlatform()` does not have them.

This new matching algorithm excludes candidates from matching with the
platform if there are other candidates with fewer total tags.  This
results in a "simplest match first" behavior, which better represents
the intent of tags in the first place.

Author: staticfloat

base/binaryplatforms.jl

@@ -1067,12 +1067,20 @@ function select_platform(download_info::Dict, platform::AbstractPlatform = HostP
         return nothing
     end
 
-    # At this point, we may have multiple possibilities.  E.g. if, in the future,
-    # Julia can be built without a direct dependency on libgfortran, we may match
-    # multiple tarballs that vary only within their libgfortran ABI.  To narrow it
-    # down, we just sort by triplet, then pick the last one.  This has the effect
-    # of generally choosing the latest release (e.g. a `libgfortran5` tarball
-    # rather than a `libgfortran3` tarball)
+    # At this point, we may have multiple possibilities.  We now engage a multi-
+    # stage selection algorithm, where we first choose simpler matches over more
+    # complex matches.  We define a simpler match as one that has fewer tags
+    # overall.  As these candidate matches have already been filtered to match
+    # the given platform, the only other tags that exist are ones that are in
+    # addition to the tags declared by the platform. Hence, selecting the
+    # minimum in number of tags is equivalent to selecting the closest match.
+    min_tag_count = minimum(length(tags(p)) for p in ps)
+    filter!(p -> length(tags(p)) == min_tag_count, ps)
+
+    # Now we _still_ may continue to have multiple matches, so we now simply sort
+    # the candidate matches by their triplets and take the last one, so as to
+    # generally choose the latest release (e.g. a `libgfortran5` tarball over a
+    # `libgfortran3` tarball).
     p = last(sort(ps, by = p -> triplet(p)))
     return download_info[p]
 end

test/binaryplatforms.jl

@@ -315,8 +315,9 @@ end
         P("x86_64", "linux"; libgfortran_version=v"5") => "linux8",
 
         # Ambiguity test
-        P("aarch64", "linux"; libgfortran_version=v"3") => "linux4",
+        P("aarch64", "linux"; libgfortran_version=v"3") => "linux3",
         P("aarch64", "linux"; libgfortran_version=v"3", libstdcxx_version=v"3.4.18") => "linux5",
+        P("aarch64", "linux"; libgfortran_version=v"3", libstdcxx_version=v"3.4.18", foo="bar") => "linux9",
 
         # OS test
         P("x86_64", "macos"; libgfortran_version=v"3") => "mac4",
@@ -327,8 +328,10 @@ end
     @test select_platform(platforms, P("x86_64", "linux"; libgfortran_version=v"4")) == "linux7"
 
     # Ambiguity test
-    @test select_platform(platforms, P("aarch64", "linux")) == "linux5"
-    @test select_platform(platforms, P("aarch64", "linux"; libgfortran_version=v"3")) == "linux5"
+    @test select_platform(platforms, P("aarch64", "linux")) == "linux3"
+    @test select_platform(platforms, P("aarch64", "linux"; libgfortran_version=v"3")) == "linux3"
+    # This one may be surprising, but we still match `linux3`, and since linux3 is shorter, we choose it.
+    @test select_platform(platforms, P("aarch64", "linux"; libgfortran_version=v"3", libstdcxx_version=v"3.4.18")) === "linux3"
     @test select_platform(platforms, P("aarch64", "linux"; libgfortran_version=v"4")) === nothing
 
     @test select_platform(platforms, P("x86_64", "macos")) == "mac4"
@@ -339,6 +342,14 @@ end
 
     # Sorry, Alex. ;)
     @test select_platform(platforms, P("x86_64", "freebsd")) === nothing
+
+    # The new "prefer shortest matching" algorithm is meant to be used to resolve ambiguities such as the following:
+    platforms = Dict(
+        # Typical binning test
+        P("x86_64", "linux") => "good",
+        P("x86_64", "linux"; sanitize="memory") => "bad",
+    )
+    @test select_platform(platforms, P("x86_64", "linux")) == "good"
 end
 
 @testset "Custom comparators" begin